14 very fev/ exceptions the above mentioned variation seems to be chiefly- 

 confined to the plants forming the genus Drosanthemum, and among them 

 may possibly be due to hydridisation, and that among the many hun- 

 dreds of fruit-capsules examined the presence or absence of s mar- 

 ginal wing or flan to the expanding-keels does not seem to vary 

 among ^lents having the same vegetative type, and is therefore a 

 more important and trustworthy character than the absence or pres- 

 ence of cell-wings or a placental tubercle. ■'-'herefore, it is evi- 

 dent thst taken alone and unaccompcnied by some floral or very defin- 

 ite vegetative character, the full development or Rudimentary nature 

 of the cell-wings, or the presence or absence of a placental tuber- 

 cle, cannot always be regarded as of generic importrnce. If, hov/- 

 ever, consistently accompanied by a floral or very pronounced vege- 

 tative character, then I consider that they should rank as of gen- 

 eric value. Because, in a vast and very complex group such as this 

 is, v/here the affinities of species are often very indefinite, it is 

 of the first importance to the botanist to have some definite struc- 

 tural character or characters whereby the generic affinity of a 

 plant may with certainty be recognised, and as during the very ela- 

 broate examination I have made it has been noticed that whenever the 

 vegetative characters are of a very pronounced type, they are usually 

 accompanied by some difference in floral or fruit structure, great 

 prominence is therefore given to vegetative characters for the pur- 

 pose of distinguishing the genera in the key to them hereafter to 

 follow, 



N. S. Brovm 

 (To be continued.) 



MSSElvlBRYANTHElvIQM. 

 Gard. Chron. HI. 8?: 32. 1930. 

 (Continued from page 14. ) 



32 I do not agree with Schwantes, however, in founding genera up- 

 on vegetative characters alone, as in the case of S^erlanzia, which 

 is identical with I-esembryanthemum in fruit and floral characters, 

 merely differing in being spiny, and one species (M, triticif orme, 

 L. 3ol.) is sniny or sninelnss. And ^scularia , Echinus, Trichodia- 

 dema and Drosenthemmn, which all differ from I^^esembryanthemum by 

 the presence of ma.rginel v/ings to the expanding keels, as well as 

 in vegetative characters, have scarcely any structural characters 

 that c-n be set down in v/ords to distinguish them for one another, 

 on account of the same characters being represented in the variable 

 genus Dposanthemum, but es their vegetative characters are so strik- 

 ingly distinct from each other and from the vegetative characters of 

 other genera, I think it of more convenience to retain them as four 

 genera than as four sections of one genus, and therefore propose to 

 retain and characterise them. But ^"^uschia , even if limited to the 

 type species H. ruDicola, Schvrant. (M. rupicolum, ^ngi.) and its 

 allies, only differs from ^-^sembryanthemum by having a tooth or 

 teeth on the keel of the leaves, for there is no structural differ- 

 ence in either flov/ers or fruits, while in the reainder of the spe- 

 cies placed under Huschia in the foliage is not toothed, although 

 variable in form, and leaf-characters alone, without some structural 

 difference, ere admittedly invalid for generic separation. Of 

 vegetative variation in the same genus we have good examples in 

 Grassula, Acacia, Surhorbia and Veronica, where it is more diverse 



